• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.3
• /
• pp.267-274
• /
• 2009

Recently, UAVs (Unmanned Aerial Vehicles) are expected much as the Unmanned Systems for various missions. These missions are often based on the Vision System. Especially, missions such as surveillance and pursuit have a process which is carried on through the transmitted vision data from the UAV. In case of small UAVs, monocular vision is often used to consider weights and expenses. Research of missions performance using the monocular vision is continued but, actually, ground and target model have difference in distance from the UAV. So, 3D distance measurement is still incorrect. In this study, Mean-Shift Algorithm, Optical Flow and Subspace Method are posed to estimate the relative depth. Mean-Shift Algorithm is used for target tracking and determining Region of Interest (ROI). Optical Flow includes image motion information using pixel intensity. After that, Subspace Method computes the translation and rotation of image and estimates the relative depth. Finally, we present the results of this study using images obtained from the UAV experiments.

• Structural Engineering and Mechanics
• /
• v.73 no.5
• /
• pp.529-542
• /
• 2020

As limited well-documented experimental data are available for assessing the attributes of different deformation components of flanged walls, few appropriate models have been established for predicting the inelastic responses of flanged walls, especially those of asymmetrical flanged walls. This study presents the experimental results for three large-scale T-shaped reinforced concrete walls and examines the variations in the flexural, shear, and sliding components of deformation with the total deformation over the entire loading process. Based on the observed deformation behavior, a simple model based on moment-curvature analysis is established to estimate flexural deformations, in which the changes in plastic hinge length are considered and the deformations due to strain penetration are modeled individually. Based on the similar gross shapes of the curvature and shear strain distributions over the wall height, a proportional relationship is established between shear displacement and flexural rotation. By integrating the deformations due to flexure, shear, and strain penetration, a new load-deformation analytical model is proposed for flexure-dominant flanged walls. The proposed model provides engineers with a simple, accurate modeling tool appropriate for routine design work that can be applied to flexural walls with arbitrary sections and is capable of determining displacements at any position over the wall height. By further simplifying the analytical model, a simple procedure for estimating the ultimate displacement capacity of flanged walls is proposed, which will be valuable for performance-based seismic designs and seismic capacity evaluations.

• Journal of Broadcast Engineering
• /
• v.24 no.4
• /
• pp.580-591
• /
• 2019

This paper focuses on to improving object detection performance using the camera and LiDAR on autonomous vehicle platforms by fusing detected objects from individual sensors through a late fusion approach. In the case of object detection using camera sensor, YOLOv3 model was employed as a one-stage detection process. Furthermore, the distance estimation of the detected objects is based on the formulations of Perspective matrix. On the other hand, the object detection using LiDAR is based on K-means clustering method. The camera and LiDAR calibration was carried out by PnP-Ransac in order to calculate the rotation and translation matrix between two sensors. For Sensor fusion, intersection over union(IoU) on the image plane with respective to the distance and angle on world coordinate were estimated. Additionally, all the three attributes i.e; IoU, distance and angle were fused using logistic regression. The performance evaluation in the sensor fusion scenario has shown an effective 5% improvement in object detection performance compared to the usage of single sensor.

• Journal of IKEEE
• /
• v.23 no.3
• /
• pp.844-851
• /
• 2019

As wearable technology is becoming more common and a part of our lives, there have been many efforts to offer various smart services with wearable devices, such as motion recognition, safety of driving, and so on. In this paper, we present a method that exploits the 9-axis inertial sensors embedded in a smartwatch to identify whether the user is a vehicle driver or not and to estimate the steering wheel turning direction in the vehicle. The system consists of three components: (i) position recognition, (ii) driver recognition, and (iii) steering-wheel turning detection components. We have developed a prototype system for detecting user's motion with Arduino boards and IMU sensors. Our experiments show high accuracy in recognizing the driver and in estimating the wheel rotation angle. The average experimental error was $11.77^{\circ}$ which is small enough to perceiver the turning direction of steering-wheel.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.13 no.1
• /
• pp.102-114
• /
• 2021

Biofouling represents an important problem in the shipping industry since it causes the increase in surface roughness. The most of ships in the current world fleet do not have good coating condition which represents an important problem due to strict rules regarding ship energy efficiency. Therefore, the importance of the control and management of the hull and propeller fouling is highlighted by the International Maritime Organization and the maintenance schedule optimization became valuable energy saving measure. For adequate implementation of this measure, the accurate prediction of the effects of biofouling on the hydrodynamic characteristics is required. Although computational fluid dynamics approach, based on the modified wall function approach, has imposed itself as one of the most promising tools for this prediction, it requires significant computational time. However, during the maintenance schedule optimization, it is important to rapidly predict the effect of biofouling on the ship hydrodynamic performance. In this paper, the effect of biofilm on the ship hydrodynamic performance is studied using the proposed performance prediction method for three merchant ships. The applicability of this method in the assessment of the effect of biofilm on the ship hydrodynamic performance is demonstrated by comparison of the obtained results using the proposed performance prediction method and computational fluid dynamics approach. The comparison has shown that the highest relative deviation is lower than 4.2% for all propulsion characteristics, lower than 1.5% for propeller rotation rate and lower than 5.2% for delivered power. Thus, a practical tool for the estimation of the effect of biofouling with lower fouling severity on the ship hydrodynamic performance is developed.

• Journal of IKEEE
• /
• v.26 no.2
• /
• pp.150-159
• /
• 2022

In this paper, we propose a physics simulation for augmented reality (AR) billiards content. The characteristics of the physics simulation for the proposed AR billiards content are as follows. First, physical equations are derived by calculating the force and moment of inertia applied to the billiards ball to realize the motion of the billiards ball similar to the real one in the AR environment. Then, we determine the velocity and angular velocity of the virtual billiards ball associated with the rotation of the virtual billiards ball with respect to the impact point. Second, using some vectors such as incidnet vector, normal vector, reflection vector, the trajectory of the virtual billiards ball would be implement. these equations are applied to AR environment so that AR billiards content could be implement. This physics simulation allows users to feel like the real world using a virtual pool table and induce them to interact with the real environment. As a result of the experiment, the accuracy range between the path of the real billiards ball and the path of the virtual billiards ball was calculated to be 97.75% to 99.11%. Therefore, it was determined that the performance of the physics simulation for the AR billiards content proposed in this paper performs similarly to the path of the real billiards ball.

• Progress in Medical Physics
• /
• v.32 no.4
• /
• pp.165-171
• /
• 2021

Purpose: This study aimed to evaluate the effect of collimator width on effective atomic number (EAN), relative electron density (RED), and stopping power ratio (SPR) measured by dual-layer dual-energy computed tomography (DL-DECT). Methods: CIRS electron density calibration phantoms with two different arrangements of material plugs were scanned by DL-DECT with two different collimator widths. The first phantom included two dense bone plugs, while the second excluded dense bone plugs. The collimator widths selected were 64 mm×0.625 mm for wider collimators and 16 mm×0.625 mm for narrow collimators. The scanning parameters were 120 kVp, 0.33 second gantry rotation, 3 mm slice thickness, B reconstruction filter, and spectral level 4. An image analysis portal system provided by a computed tomography (CT) manufacturer was used to derive the EAN and RED of the phantoms from the combination of low energy and high energy CT images. The EAN and RED were compared between the images scanned using the two different collimation widths. Results: The CT images with the wider collimation width generated more severe artifacts, particularly with high-density material (i.e., dense bone). RED and EAN for tissues (excluding lung and bones) with the wider collimation width showed significant relative differences compared to the theoretical value (4.5% for RED and 20.6% for EAN), while those with the narrow collimation width were closer to the theoretical value of each material (2.2% for EAN and 2.3% for RED). Scanning with narrow collimation width increased the accuracy of SPR estimation even with high-density bone plugs in the phantom. Conclusions: The effect of CT collimation width on EAN, RED, and SPR measured by DL-DECT was evaluated. In order to improve the accuracy of the measured EAN, RED, and SPR by DL-DECT, CT scanning should be performed using narrow collimation widths.

• Composites Research
• /
• v.36 no.4
• /
• pp.259-263
• /
• 2023

This study investigated the open hole tensile (OHT) properties of carbon fiber/epoxy composites and compared them to bolt joint tensile (BJT) properties. The net nominal modulus and strength (1376 MPa) were found to be higher than the gross nominal strength (1041 MPa), likely due to increasing hole size. The OHT and BJT specimens exhibited similar stiffness, as expected without bolt rotation causing secondary bending. OHT specimens experienced a sharp drop in stress indicating unstable crack propagation, delamination, and catastrophic failure. BJT specimens failed through shear out on the bolt side and bearing failure on the nut side, involving fiber kinking, matrix splitting, and delamination, resulting in lower strength compared to OHT specimens. The strength retention of carbon fiber/epoxy composites with open holes was 66%. Delamination initiation at the hole's edge caused a reduction in the stress concentration factor. Filling the hole with a bolt suppressed this relieving mechanism, leading to lower strength in BJT specimens compared to OHT specimens. Bolt joint efficiency was calculated as 15%. The reduction in strength in bolted joints was attributed to fiber-matrix splitting and delamination, aligning with Hart Smith's bolted joint efficiency diagram. These findings contribute to materials selection and structural reliability estimation for carbon fiber/epoxy composites. They highlight the behavior of open hole and bolt joint configurations under tensile loading, providing valuable insights for engineering applications.

• Journal of KIISE:Software and Applications
• /
• v.35 no.10
• /
• pp.600-607
• /
• 2008

This paper proposes a robust human detection method irrespective of their pose variation using the multiple oriented 2D elliptical filters (MO2DEFs). The MO2DEFs can detect the humans regardless of their poses unlike existing object oriented scale adaptive filter (OOSAF). To overcome OOSAF's limitation, we introduce the MO2DEFs whose shapes look like the oriented ellipses. We perform human detection by applying four different 2D elliptical filters with specific orientations to the 2D spatial-depth histogram and then by taking the thresholds over the filtered histograms. In addition, we determine the human pose by using convolution results which are computed by using the MO2DEFs. We verify the human candidates by either detecting the face or matching head-shoulder shapes over the estimated rotation. The experimental results showed that the accuracy of pose angle estimation was about 88%, the human detection using the MO2DEFs outperformed that of using the OOSAF by $15{\sim}20%$ especially in case of the posed human.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.2
• /
• pp.79-88
• /
• 2004

Gaze detection is to locate the position on a monitor screen where a user is looking by computer vision. Gaze detection systems have numerous fields of application. They are applicable to the man-machine interface for helping the handicapped to use computers and the view control in three dimensional simulation programs. In our work, we implement it with a computer vision system setting a IR-LED based single camera. To detect the gaze position, we locate facial features, which is effectively performed with IR-LED based camera and SVM(Support Vector Machine). When a user gazes at a position of monitor, we can compute the 3D positions of those features based on 3D rotation and translation estimation and affine transform. Finally, the gaze position by the facial movements is computed from the normal vector of the plane determined by those computed 3D positions of features. In addition, we use a trained neural network to detect the gaze position by eye's movement. As experimental results, we can obtain the facial and eye gaze position on a monitor and the gaze position accuracy between the computed positions and the real ones is about 4.8 cm of RMS error.